A method, device, medium and equipment for diagnosing diesel engine fault with sample imbalance

Abstract

The application discloses a diesel engine fault diagnosis method for sample imbalance, comprising: obtaining a target vibration signal of a diesel engine, wherein the target vibration signal comprises a target vibration signal under a normal state and target vibration signals under different fault states; performing wavelet packet noise compressed sensing sample enhancement processing on the target vibration signal to obtain an enhanced signal sample; performing continuous wavelet transform on the enhanced signal sample to obtain a time-frequency graph; performing gray scaling on the time-frequency graph according to an edge detection algorithm, calculating an edge density, and constructing an edge density vector according to the state of the target vibration signal; constructing an edge density threshold value of different states based on the edge density vector, calculating the edge density of a to-be-detected signal time-frequency graph, and comparing the edge density with the edge density threshold value to obtain the state of the diesel engine. The application improves the sample balance and fault diagnosis capability of the marine diesel engine, and has important significance for ensuring safe navigation of a ship, reducing operation cost and prolonging the service life of equipment.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of diesel engine diagnosis, and particularly relates to a sample imbalance diesel engine fault diagnosis method, device, medium and equipment. BACKGROUND

[0002] Marine diesel engine as the core equipment of the ship power system, plays a vital role in the shipping industry. It is widely used in various merchant ships, tankers, cargo ships and other large ships, providing strong propulsion and stable power support. However, marine diesel engines are long-term operated in harsh marine environments, facing high load, high temperature, seawater corrosion and other adverse factors, prone to various faults, and it is necessary to diagnose the faults of marine diesel engines. However, the fault diagnosis method of marine diesel engine needs to use a large amount of data to adjust the model, and the diesel engine is in normal state for a long time, so the data of fault state is much less than that of normal state, causing the problem of unbalanced data samples, which leads to inaccurate construction of fault model. It is difficult to find faults in time, which affects the efficiency of navigation, and serious accidents may even cause serious damage to equipment and safety accidents. SUMMARY

[0003] The present application provides a sample imbalance diesel engine fault diagnosis method, device, medium and equipment to solve the above problems existing in the prior art.

[0004] To achieve the above purpose, the present application provides a sample imbalance diesel engine fault diagnosis method, which comprises the following steps:

[0005] Obtaining the target vibration signal of the diesel engine, wherein the target vibration signal includes the target vibration signal under normal state and the target vibration signal under different fault states;

[0006] Performing wavelet packet noise compressed sensing sample enhancement processing on the target vibration signal to obtain enhanced signal samples;

[0007] Performing continuous wavelet transform on the enhanced signal samples to obtain a time-frequency graph;

[0008] According to the edge detection algorithm, the time-frequency graph is grayed, and the edge density is calculated, and the edge density vector is constructed according to the state of the target vibration signal;

[0009] Based on the edge density vector, the edge density threshold of different states is constructed, the edge density of the to-be-detected signal time-frequency graph is calculated and compared with the edge density threshold, and the state of the diesel engine is obtained.

[0010] Preferably, the wavelet packet noise compressed sensing sample enhancement includes:

[0011] Wavelet packet decomposition is performed on the target vibration signal to obtain a plurality of sub-signals;

[0012] One of the sub-signals is encoded by compressive sensing encoding, and a Gaussian mixed noise is constructed and superimposed with the encoded sub-signal;

[0013] The superimposed signal is decoded by compressive sensing decoding to obtain an enhanced component;

[0014] The enhanced component and other non-enhanced components are reconstructed.

[0015] Preferably, the wavelet packet decomposition expression is:

[0016]

[0017] In the formula, x(t) represents a signal to be decomposed, J is the number of decomposition layers, 2 j is the number of sub-signals of each layer, d j,k is the kth sub-signal of the jth layer, φ j,k is the kth wavelet packet basis function of the jth layer.

[0018] Preferably, the encoding operation by compressive sensing encoding is expressed as:

[0019] y=Φd j,k ;

[0020] In the formula, Φ represents a measurement matrix, d j,k is the kth sub-signal of the jth layer.

[0021] Preferably, the expression of the Gaussian mixed noise is:

[0022]

[0023] In the formula, g represents the generated Gaussian mixed noise, z represents the number of Gaussian distributions, w i represents the weight of the ith Gaussian distribution, η represents the generated Gaussian noise according to μ i and σ i , μ i represents the mean of the ith Gaussian distribution, σ i represents the standard deviation of the ith Gaussian distribution.

[0024] Preferably, the expression of the continuous wavelet transform is:

[0025]

[0026] In the formula, x(t) is a one-dimensional time-domain signal, Ψ(*) is a wavelet basis function, t represents time, a is a scale parameter, and b is a transform form of a translation parameter.

[0027] Preferably, the edge density calculation comprises:

[0028] The time-frequency diagram is grayed by the brightness method, and Gaussian filtering is performed to remove image noise;

[0029] The image gradient is calculated by the Sobel operator to obtain the image edge;

[0030] The edge pixel ratio is calculated to obtain the edge density.

[0031] The application also provides a computer device comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to realize the steps of the method.

[0032] The application also provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to realize the steps of the method.

[0033] The application also provides an electronic device comprising a memory and a processor, wherein the memory is configured to store a program, and the processor is configured to execute the program to realize the steps of the method.

[0034] Compared with the prior art, the application has the following advantages and technical effects:

[0035] The application discloses a diesel engine fault diagnosis method for sample imbalance, and comprises the following steps: obtaining a target vibration signal of a diesel engine, wherein the target vibration signal comprises a target vibration signal in a normal state and target vibration signals in different fault states; performing wavelet packet noise compression sensing sample enhancement processing on the target vibration signal to obtain an enhanced signal sample; performing continuous wavelet transform on the enhanced signal sample to obtain a time-frequency diagram; performing gray processing on the time-frequency diagram according to an edge detection algorithm, and calculating an edge density, and constructing an edge density vector according to the state of the target vibration signal; constructing an edge density threshold of different states based on the edge density vector, calculating the edge density of a time-frequency diagram of a to-be-detected signal, and comparing the edge density with the edge density threshold to obtain a state of the diesel engine. The application improves the sample balance and fault diagnosis capability of the marine diesel engine, and has important significance for ensuring safe navigation of a ship, reducing operation cost and prolonging the service life of equipment. BRIEF DESCRIPTION OF DRAWINGS

[0036] The accompanying drawings, which form a part of this application, are included to provide a further understanding of the application and are incorporated in and constitute a part of this application. The embodiments of this application and their explanations are used to explain this application, and do not constitute an improper limitation on this application. In the drawings:

[0037] Figure 1 The method flowchart of the embodiments of the application;

[0038] Figure 2 Fig. 1 is a normal sample image according to an embodiment of the present application, wherein (a) is an enhanced effect image of the normal sample, and (b) is an original image of the normal sample;

[0039] Figure 3 Fig. 2 is an abnormal sample image according to an embodiment of the present application, wherein (a) is an enhanced effect image of the abnormal sample, and (b) is an original image of the abnormal sample. DETAILED DESCRIPTION

[0040] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0041] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown herein.

[0042] The technologies involved will be introduced below:

[0043] Wavelet packet decomposition is a technique for multi-scale and multi-frequency analysis of signals. It recursively decomposes a signal into sub-signals containing different frequency components, allowing for in-depth exploration of local features of the signal. This decomposition method is very flexible, allowing users to select different wavelet basis functions and decomposition depths according to their needs, and is particularly suitable for analyzing nonlinear and non-stationary complex signals. Wavelet packet decomposition has wide applications in signal processing, image analysis, audio analysis, and biomedical signal analysis, etc. It provides detailed frequency and time information of signals, helping us to more accurately understand and process various signals.

[0044] Continuous wavelet transform (CWT) is a powerful signal analysis tool that can reveal local features of a signal such as discontinuities, abrupt changes, and spikes by analyzing the signal at different scales and positions. This transform uses wavelet functions to probe the signal, similar to using magnifying glasses of different magnifications to observe the details of the signal. CWT is particularly suitable for analyzing non-stationary signals, i.e. signals whose statistical properties change over time. It can adapt to the nonlinearity and non-stationarity of the signal, and by selecting appropriate wavelet basis functions, it increases the flexibility of analysis. Continuous wavelet transform has wide applications in signal processing, image analysis, seismology, and financial analysis, etc. It is a very valuable analysis tool because it can provide time-frequency information of signals.

[0045] The Sobel operator identifies edges by calculating the gradient magnitude at each pixel in an image. This operator uses two 3x3 convolutional kernels: one to detect horizontal brightness changes and the other to detect vertical brightness changes. Using the gradients in these two directions, the Sobel operator can determine the intensity and direction of the edge. This method is particularly effective for horizontal and vertical edges in images, but its ability to detect diagonal edges is weaker. The Sobel operator is widely used in practical applications due to its simplicity and effectiveness, especially when fast and relatively accurate edge detection is required.

[0046] Example 1

[0047] like Figure 1 As shown, this embodiment provides a method for diagnosing diesel engine faults due to sample imbalance, including the following steps:

[0048] Acquire the target vibration signal of the diesel engine, which includes the target vibration signal under normal conditions and the target vibration signal under different fault conditions;

[0049] The target vibration signal is subjected to wavelet packet noise compression sensing sample enhancement processing to obtain the enhanced signal sample;

[0050] Continuous wavelet transform is performed on the enhanced signal samples to obtain the time-frequency diagram;

[0051] The time-frequency image is converted to grayscale based on the edge detection algorithm, and the edge density is calculated. An edge density vector is constructed based on the state of the target vibration signal.

[0052] Edge density thresholds for different states are constructed based on edge density vectors. The edge density of the time-frequency map of the signal to be detected is calculated and compared with the edge density thresholds to obtain the state of the diesel engine.

[0053] The specific steps are as follows:

[0054] 1. Wavelet packet noise compression sensing sample enhancement:

[0055] The vibration signals S collected by the acceleration sensor in different states are unbalanced because the time of normal state is much longer than that of abnormal state, resulting in that the number of abnormal state samples is much less than that of normal state samples. The imbalance of samples may lead to distortion of evaluation index of data, and the evaluation index is affected by the imbalance of samples and thus is biased to the majority type, thereby affecting the accuracy. To solve the problem of sample imbalance, a wavelet packet noise compressed sensing sample enhancement method is used to enhance the data samples, so as to obtain balanced samples. The specific implementation method is as follows: a sample signal is decomposed into n sub-signals by wavelet packet decomposition, one sub-signal is selected, compressed sensing encoding is used to encode the data, a Gaussian mixed noise is constructed and superimposed with the encoded signal, the superimposed signal is decoded by compressed sensing, and an enhanced component is obtained. Finally, the component is reconstructed with other components to obtain a signal similar to but not the same as the original signal. The above operation is repeated n times to complete the enhancement of a sample signal. The implementation method of wavelet packet decomposition is shown in formula (1).

[0056]

[0057] In formula (1), x(t) represents the signal to be decomposed, J represents the decomposition layer, 2 j is the number of sub-signals of each layer, d j,k is the kth sub-signal of the jth layer. φ j,k is the kth wavelet packet basis function of the jth layer, which is the form obtained by scaling and shifting the mother wavelet. A plurality of sub-signals can be obtained by formula (1), and one of the sub-signals d j,k is selected and compressed sensing encoding is performed. j,k Compressed sensing considers that the signal d j,k is sparse, and the sub-signal d j,k is compressed and sampled by a measurement matrix Φ, as shown in formula (2):

[0058] y = Φd j,k (2)

[0059] In the above process, a measurement value y is generated by using a randomly generated matrix Φ for sampling, and the signal is encoded. The sub-signal d j,k is effectively reconstructed. On the basis of the encoding, to realize the differentiation of the sample enhancement signal, a Gaussian mixed noise (with different standard deviations) is added to the signal, and the mathematical expression of the Gaussian mixed noise is shown in formula (4). To ensure the consistency of the signal quality after adding the noise, the signal standard deviation of the Gaussian mixed noise needs to be calculated, as shown in formula (3).

[0060]

[0061] where σ represents the calculated standard deviation of Gaussian noise, N represents the length of signal d j,k g represents the generated Gaussian noise, z represents the number of Gaussian distributions, w i represents the weight of the i-th Gaussian distribution, μ i represents the mean of the i-th Gaussian distribution, σ i represents the standard deviation of the i-th Gaussian distribution. A new signal y noisy is obtained by superimposing the compressed sensing encoded signal with the Gaussian mixed noise. On this basis, the compressed sensing decoding operation is performed on the signal, and the purpose of the decoding operation is to obtain similar but different sub-signals. The process is to reconstruct the original signal x through an optimization problem, and the optimization process is shown in equation (5)

[0062]

[0063] The optimization objective of this process is to find an x that minimizes equation (5), where λ is a regularization parameter used to control sparsity, ||·||2 represents the L2 norm, i.e. the square root of the sum of squares. ||·||1 represents the L1 norm, i.e. the sum of absolute values.

[0064] According to equation (1), the processed sub-signals are replaced with the atomic signals, and the signal is reconstructed to obtain a signal sample similar to but not identical to the original sample, thereby achieving enhancement of the fault sample. Figure 2 With Figure 3 the schematic diagram of normal samples and abnormal samples generated by the sample enhancement method.

[0065] 2. Time-frequency plot drawing method:

[0066] Continuous wavelet transform is an effective tool for analyzing signals in time and frequency domains simultaneously, with the advantages of time-frequency localization and multi-scale analysis. It mainly performs continuous inner product on one-dimensional signals through equation (6) using wavelet basis functions of different scales. In order to further extract fault feature information from the signal, continuous wavelet transform is used to generate time-frequency plots.

[0067]

[0068] where y(t) is a one-dimensional time-domain signal, Ψ(*) is a wavelet basis function, a is a scale parameter, and b is a translation parameter. In the continuous wavelet time-frequency plot, the energy distribution in different frequency ranges can reflect different fault characteristics, so the continuous wavelet time-frequency plot can extract fault characteristics by analyzing the energy distribution in different frequency ranges and the energy concentration area.

[0069] 3. Simple edge detection diagnosis method:

[0070] In two-dimensional images, by identifying the regions of significant changes in brightness in the image, these regions correspond to important features in the image, for which the edge detection can be used for fault diagnosis of two-dimensional images. To achieve edge detection, it is necessary to first perform image graying, which can be achieved by brightness method, as shown in equation (7).

[0071] I gray (x,y)=w R ×R(x,y)+w G ×G(x,y)+w B ×B(x,y) (7)

[0072] where I gray is the gray image, R(x,y), G(x,y), I(x,y) are the pixel values of red, green and blue respectively. w R , w G , w B are three different weights. On this basis, using Gaussian filter, smooth the image and remove noise, reduce the interference factors in the detection process, the specific method is shown in equation (8).

[0073]

[0074] where p represents the standard deviation of the Gaussian kernel, control the degree of smoothing, through the above process to get the smoothed image. In the signal after smoothing using Sobel operator for image gradient calculation, the gradient of the image in x and y direction. Gradient calculation process is shown in equation (9) and (10), where the size and direction of the gradient calculation is shown in equation (11) and (12).

[0075]

[0076] θ(x,y)=atan2(G y (x,y),G x (x,y)) (12)

[0077] where G x is the horizontal gradient, G y is the vertical gradient, the gradient represents the edge strength of a point in the image, θ is the gradient angle. To simplify the calculation process, the gradient direction is quantized as 0°, 45°, 90°, 135°. In different gradient directions, if the gradient value of the current pixel is not the local maximum value along the gradient direction, it is set to 0, if it is, the pixel point is retained, so as to refine the edge. Using double threshold detection, three different types of strong edge, weak edge, non-edge are distinguished, by setting high threshold T high and low threshold T low . The mathematical expression of double threshold detection is shown in equation (13).

[0078]

[0079] After the above processing, in order to measure the edge information density in the image signal and thus determine the number of fault components, the edge density is used to calculate the proportion of edge pixels, and a specific implementation formula is shown in formula (14).

[0080]

[0081] where W and H are the width and height of the image respectively, and WxH is equal to the total number of pixels in the image. The edge density information of the fault time-frequency diagram in different states is calculated by the above method, and the standard deviation σ edge and the mean μ edge are calculated. According to the Gaussian distribution principle, 99.7% of the data is within the range of μ±3σ. Therefore, the Gaussian distribution principle can be used to calculate the mean σ edge and the standard deviation μ edge of the edge density of multiple images in different states. Different state thresholds are constructed, as shown in formula (15).

[0082]

[0083] The embodiment improves the sample balance and fault diagnosis capability of the marine diesel engine, which is of great significance to ensure the safe navigation of the ship, reduce the operating cost, and prolong the service life of the equipment.

[0084] The embodiment also provides a computer device, which includes a memory, a processor, and a computer program stored in the memory, and the processor executes the computer program to realize the steps of the method.

[0085] The embodiment also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the steps of the method.

[0086] The embodiment also provides an electronic device, which includes a memory and a processor, the memory is used for storing a program, and the processor is used for executing the program to realize the steps of the method.

[0087] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited to this. Any changes or replacements within the technical range disclosed in the present application can be easily thought of by those skilled in the art, and should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method of diesel engine fault diagnosis with sample imbalance, characterized by, The method comprises the following steps: obtaining a target vibration signal of a diesel engine, wherein the target vibration signal comprises a target vibration signal in a normal state and target vibration signals in different fault states; performing wavelet packet noise compressed sensing sample enhancement processing on the target vibration signal to obtain an enhanced signal sample; the wavelet packet noise compressed sensing sample enhancement comprises: performing wavelet packet decomposition on the target vibration signal to obtain a plurality of sub-signals; performing encoding operation on a sub-signal through compressed sensing encoding, and constructing a Gaussian mixed noise to be superimposed with the encoded sub-signal; decoding the superimposed signal through compressed sensing decoding to obtain an enhanced component; reconstructing the enhanced component and other non-enhanced components; performing continuous wavelet transform on the enhanced signal sample to obtain a time-frequency graph; graying the time-frequency graph according to an edge detection algorithm, and calculating edge density, and constructing an edge density vector according to the state of the target vibration signal; constructing edge density thresholds of different states based on the edge density vector, calculating the edge density of a time-frequency graph of a to-be-detected signal, and comparing the edge density with the edge density thresholds to obtain the state of the diesel engine.

2. The method of claim 1, wherein, The wavelet packet decomposition expression is: ； In the formula, Indicates the signal to be decomposed. It is the number of decomposition layers. It is the number of sub-signals in each layer. It is the first Layer A sub-signal, It is the first Layer Wavelet packet basis functions.

3. The method of claim 1, wherein, The expression of the encoding operation through compressed sensing encoding is: ； In the formula, denotes a measurement matrix, is the first layer the first sub-signal.

4. The method of claim 1, wherein, The expression of the Gaussian mixed noise is: ； where g denotes generated Gaussian mixture noise, denotes the number of Gaussian distributions, denotes the weight of the i-th Gaussian distribution, denotes the i-th Gaussian distribution according to and generated Gaussian noise, denotes the mean of the i-th Gaussian distribution, denotes the standard deviation of the i-th Gaussian distribution.

5. The method of claim 1, wherein, The expression of the continuous wavelet transform is: ； wherein is a one-dimensional time signal, is a wavelet basis function, t denotes time, a is a scale parameter, and b is a translation parameter.

6. The method of claim 1, wherein, The calculation of the edge density comprises: graying the time-frequency graph through brightness method, and performing Gaussian filtering to remove image noise; performing image gradient calculation through a Sobel operator to obtain image edges; calculating the proportion of edge pixels to obtain the edge density.

7. A computer apparatus comprising a memory, a processor, and a computer program stored on the memory, wherein the computer program, when executed by the processor, causes the processor to perform the method of any one of claims 1 to 6. The processor executes the computer program to realize the steps of the method of any one of claims 1-6.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method of any one of claims 1-6.

9. An electronic device, comprising: comprise: a memory and a processor; the memory is used to store a program; the processor is used to execute the program to realize the steps of the method of any one of claims 1-6.

Images